A new study by Previsible reveals that AI language models (LLMs) are gaining traction as a referral traffic source and are seeing significant changes in search behavior.
After analyzing over 30 websites, we found that Perplexity and ChatGPT are emerging as alternatives to traditional search engines.
Previsible co-founder David Bell believes this is a sign that Google’s growth is stalling.
“Google has basically plateaued and its search dominance is starting to decline. The reason is that people are using different tools like ChatGPT, Claude, Co-pilot, and Bing to better resolve search intent. Because they’re starting to use the experience.”
Here are some key takeaways from the research. It does not provide the complete picture, but it does provide the best information available at this time.
Main findings
referral traffic
The study found that Perplexity and ChatGPT accounted for approximately 37% of LLM referral traffic, followed by CoPilot and Gemini at 12-14% each.
In particular, the finance sector accounts for the majority of LLM-driven traffic, accounting for 84% of all referrals analyzed.
In a video walkthrough of the study, Bell explains:
“We are seeing a significant increase in traffic from language models, especially in finance, as Perplexity and other language models have integrations or relationships with various platforms that allow more direct access to users. It could be because you have it.”
Content distribution
The study revealed that blog posts receive 77.35% of LLM referral traffic, followed by:
Homepage visit (9.04%) News content (8.23%) Guide (2.35%)
“In the age of AI search, information content remains important,” notes Bell. However, language models don’t feature product pages prominently, so we advise them to focus on conversion rate optimization (CRO) and user journeys.
According to this study, product pages receive less than 0.5% of LLM referral traffic, suggesting challenges for e-commerce strategies.
Looking to the future
LLM referral traffic currently represents 0.25% of total traffic in the most affected sectors, but the study points to a significant increase.
In the past 90 days of research, Previsible has discovered:
900% increase in ChatGPT referrals for the event industry ChatGPT traffic for e-commerce and finance increased by over 400% Consistent growth across all models except CoPilot
Bell explains what this means for websites:
“If we extrapolate and average all of this out, we would expect organic or AI traffic to increase by approximately 200% every 90 days over the next 12 months, which could be up to 20% of all traffic to websites. There is a sex.”
🚨 Free Tool Alert 🚨
Previsible has created a free Looker Studio dashboard that allows businesses to track website traffic from language learning models (LLMs) such as ChatGPT, Perplexity, Co-pilot, Gemini, and Claude.
You can select Google Analytics 4 (GA4) accounts and data ranges to view data.
The dashboard shows:
Organic Sessions: Total organic sessions during the selected time period. Total LLM Sessions: Number of sessions from LLM and their breakdown. LLM traffic over time: A line chart showing trends in LLM traffic. A separate line is displayed for each LLM. LLM Traffic by Landing Pages: Table of top landing pages from LLM traffic, including sessions, percentage, average time on page, and comparison to site average.
How this helps
By analyzing this data, you can:
Evaluate the impact of LLM on your traffic compared to organic traffic. Find out which LLMs are driving the most traffic and adjust your content. Track your LLM traffic growth over time and adapt your strategy. Discover popular landing pages among LLM users and improve them to increase engagement. Compare the time LLM users spend on each page to your site average to identify areas for improvement.
In summary
There are three key takeaways from this study:
Financial websites have the highest LLM referral activity, with blog content accounting for the majority of visits Product pages rarely appear in LLM results, suggesting the need for a tailored e-commerce strategy The growth rate is significant and could reach 20% of total traffic within a year if current trends continue
When looking at these trends, it’s important to maintain a balanced approach to acquiring traffic and optimizing your strategy.
Don’t pursue AI traffic if it could negatively impact your sales.
AI language models are becoming a new source of website traffic. However, it currently accounts for only about 0.25% of total traffic in the most affected sectors.
It will be interesting to see how this number changes by next year.
Featured image: Koshiro/shutterstock
